The Importance of Data in Trust & Safety

Data plays a critical role in trust and safety in a variety of ways. First, data facilitates progress throughout the product development life cycles, with “product” broadly defined (including but not limited to online platforms, apps, content moderation tools, etc.).

Data aids in:

  • Understanding the size and severity of potential problems;
  • Evaluating user sentiments and challenges;
  • Informing investments and priorities within policy, product, and operations;
  • Measuring the success of decisions and investments (within product, policy, and enforcement).

For example, data can help trust and safety teams better understand how users interact with their systems so they can adjust policies accordingly. By collecting relevant data points such as customer feedback, usage trends, etc., trust and safety teams can determine if they are meeting their goals, or if changes need to be made to achieve them. Data gives trust and safety teams insight into what stakeholders want from an online experience as well as supporting evidence for setting accurate expectations. Without ongoing and transparent communication, misunderstandings or rumors about a platform’s policies may replace the facts. Data helps identify and correct discrepancies between what is being said about the platform and what actually happens within the system. 

In addition, data is instrumental in communicating and working cross functionally with both internal and external stakeholders. By providing insights into user behavior using concrete evidence and numbers, trust and safety teams demonstrate their value in a way that resonates with these stakeholders. Furthermore if there is an increase in malicious activities or threats detected by the team’s analytics platform, data can be used to illustrate the need for additional investment in trust and safety operations. 

Data-Information-Knowledge-Wisdom (DIKW) Framework

Not all numbers are the same. The value that data professionals provide is bringing structure to collections of facts and figures, extracting insights and understanding, and ultimately informing and driving decisions that have a positive impact on goals. Organizations typically use the DIKW framework in trust and safety because (1) understanding what kinds of knowledge and wisdom are desirable helps determine how data should be logged and processed; (2) identifying strengths and weaknesses of the data helps inform what knowledge and wisdom can be gained. 

In data professions, many commonly used words have definitions that are more specific or contextual, and sometimes need to be clarified. Within the DIKW framework:

  • Data is raw and unprocessed facts, stored inputs, or observations.
  • Information is data that has been organized and structured so that it can be viewed in context.
  • Knowledge is information that has been interpreted to add meaning – identifying patterns, trends, and outliers.
  • Wisdom is knowledge used to reach clear and trustworthy conclusions and to drive impactful decisions.

An example of this framework in action might be taking individual content reports (data), summarizing the daily rate of reporting (information), identifying a sudden uptick in the number of reports on a specific topic (knowledge), and discovering the cause is a new viral policy violation that should be addressed (wisdom).